Seasonal variation of serum KL-6 concentrations is greater in patients with hypersensitivity pneumonitis

Electronic medical records of patients with registered diagnoses of ILD who consulted our university hospital from April 1, 2009 to March 31, 2014 were reviewed by two pulmonologists. Patients with ILD whose serum KL-6 concentrations were measured at least four times over a period of six months and more than once in one season were selected for analysis. We excluded undiagnosed cases with ILD, ILD cases with lung cancer, drug-induced pneumonitis, radiation pneumonitis, PCP, or CMV pneumonia, each of which has been reported to increase serum KL-6 concentrations. The final study population consisted of 96 ILD patients classified as IPF (n = 16), NSIP (n = 16), CVD-IP (n = 33), House-HP (n = 9), Bird-HP (n = 9), and combined pulmonary fibrosis and emphysema (CPFE, n = 13), based on diagnostic criteria for each. This study complied with the Declaration of Helsinki. The Ethical Review Board of the Kochi Medical School, Kochi University, reviewed the study design and protocol. Approval was waived according to the Ethical Guideline for Clinical Research issued by the Ministry of Health, Labour and Welfare, Japan. Informed consent for the usage of data had been obtained from all patients at their consultation and patients who refused were not included in this study.

IPF was clinically diagnosed based on the ATS/ERS/JRS/ALAT guideline using the presence of basal honeycomb shadows on the chest high-resolution computed tomography (HRCT) (all cases), bronchoalveolar lavage (BAL) (n = 5), transbronchial lung biopsy (TBLB) (n = 5) and exclusion of other causes such as NSIP, CVD-IP, HP, and CPFE [16]. Eight patients with NSIP were histologically diagnosed based on the ATS/ERS multidisciplinary consensus classification of IIPs with surgical lung biopsy [17]. Other 8 NSIP patients were clinically diagnosed by the chest HRCT findings compatible with NSIP pattern, increased lymphocytosis in BAL fluid, lack of granuloma in TBLB specimens, and exclusion of IPF, CVD-IP, HP, or CPFE. CVD-IP was diagnosed based on symptoms, laboratory findings of CVD, and interstitial shadows on chest HRCT. HP was diagnosed by combinations of the presence of specific IgG antibodies against pigeon, budgerigar (Phadia AB, Uppsala, Sweden) for Bird-HP or Trichosporon asahii for House-HP, spontaneous remission of interstitial shadows after isolation from causative antigens, positive results on an environmental provocation test, bronchoscopic findings including BAL (n = 17), TBLB (n = 16) and/or surgical lung biopsy (n = 8) [10]. House-HP refers to HP positive for specific IgG antibody against Trichosporon asahii. BAL samples were obtained by instillation of 50 ml of saline 3 times and differential cell numbers were determined at the diagnosis of HP without treatment. HP was further classified into acute HP (acute and subacute forms) based on acute onset of symptoms and the appearance of mainly ground glass opacities on chest HRCT, or chronic HP based on gradual progression of symptoms and the presence of mainly interstitial fibrotic shadows on chest HRCT. All cases of Bird-HP in the present study were related to the use of feather products. CPFE was diagnosed by the presence of both emphysematous changes in the upper lobes and interstitial fibrotic changes in the lower lobes on chest HRCT, and exclusion of other causes such as CVD-IP, HP, or drug-induced pneumonitis [18]. Two cases of IPF, five cases of NSIP, five cases of CVD-IP, two cases of House-HP, six cases of Bird-HP, and 12 cases of CPFE were carefully observed throughout their clinical courses without treatment due to patients’ refusal of using pirfenidone, corticosteroids or immunosuppressants. Environmental improvement to avoid exposure to causative antigens was instructed to all patients at the time of their final HP diagnosis.

Serum samples were collected sequentially from patients 4 to 60 times (median 15 times) during their clinical course (median 903 days, range 195–1703 days). The frequency and duration of KL-6 measurement was not different among each group of ILD (Table 1). Serum KL-6 concentrations were measured by the chemiluminescent enzyme immunoassay LUMIPULSE KL-6 using a KL-6 antibody (EIDIA Co., Ltd, JPN). The cut-off value for serum KL-6 was set at 435 U/ml according to the manufacturer’s instructions. The maximum or minimum concentrations of serum KL-6 in each patient’s clinical course was defined as Kmax or Kmin, respectively. Similarly, serum KL-6 concentrations during the summer from June 1 to September 30 was defined as Smax or Smin, and serum KL-6 concentrations during the winter from November 1 to February 28 as Wmax or Wmin. The variation of serum KL-6 concentrations over each patient’s clinical course (Kmax-Kmin or Kmax/Kmin), during the summer (Smax-Smin or Smax/Smin), during the winter (Wmax-Wmin or Wmax/Wmin), and seasonal changes in serum KL-6 concentrations between summer and winter (Smax-Wmax, Smax/Wmax, Smin-Wmin, or Smin/Wmin) were retrospectively analyzed.

Data are presented as the median (interquartile range). Statistical analysis of data was performed using IBM SPSS Statistics version 19 for Windows (IBM Corp, NY, USA). The comparisons of serum KL-6 concentrations among ILD were tested by the Kruskal-Wallis test. Differences in serum KL-6 concentrations between two groups were evaluated by the two-tailed Mann–Whitney U-test. Serum KL-6 concentrations and KL-6 ratios were further analyzed using receiver operating characteristic (ROC) curves to determine the cut-off levels that resulted in the optimal diagnostic accuracy for the diagnosis of HP, House-HP or Bird-HP. Quantitative differences were tested by the Pearson’s chi-square test for goodness of fit or the Fisher’s exact probability test. A P-value of < 0.05 was considered statistically significant. The Holm’s sequential Bonferonni correction was used to adjust the levels of significance for multiple comparisons.

Patients with CVD-IP were significantly younger than patients with IPF (Table 1). Patients with CVD-IP (23 out of 33) and NSIP (9 out of 16) were predominantly female, and all cases with CPFE were male. Four cases with IPF and one case with NSIP had acute exacerbation during their clinical course [842 (527) days], and two cases of IPF and one case of NSIP were deceased.

The maximum KL-6 concentrations over the patients’ clinical courses (Kmax) were significantly higher in patients with NSIP compared to CPFE (Table 1, Figure 1A). The minimum KL-6 (Kmin) was significantly higher in patients with IPF than with CVD-IP. The variation of serum KL-6 concentrations over their clinical courses (Kmax-Kmin or Kmax/Kmin) was significantly greater in patients with Bird-HP than with CPFE (Table 1, Figure 1B). The variation of KL-6 during the summer (Smax/Smin) in patients with House-HP was significantly greater than Bird-HP (P = 0.003) and showed a tendency to be greater than CPFE (P = 0.004), CVD-IP (P = 0.021), IPF (P = 0.033), or NSIP (P = 0.037) (Table 1, Figure 1C). The maximum KL-6 concentrations during the winter (Wmax) and Wmax-Wmin were significantly higher in patients with Bird-HP than CVD-IP (Table 1, Figure 1D).

It is particularly noteworthy that Bird-HP and House-HP showed significantly greater seasonal changes in serum KL-6 concentrations than the other ILD. Serum KL-6 concentrations between the summer and the winter (Smax-Wmax or Smax/Wmax) were greater in patients with House-HP than Bird-HP (P < 0.001) or CPFE (P = 0.003) (Table 1, Figure 2A). Smax/Wmax in patients with Hosue-HP showed a tendency to be greater than IPF (P = 0.008) or CVD-IP (P = 0.01). Eight out of nine patients (88.9%) with Bird-HP showed greater increases in serum KL-6 concentrations in the winter compared to the summer; as such, Smax-Wmax in patients with Bird-HP showed negative values and Smax-Wmax or Smax/Wmax was significantly lower than the other ILD (Table 1, Figure 2A). These data suggest that serum KL-6 concentrations were significantly increased in the summer in patients with House-HP or in the winter in patients with Bird-HP. Typical examples of sequential seasonal changes in serum KL-6 concentrations in a patient with House-HP who had a specific IgG antibody against Trichosporon asahii and in a patient with Bird-HP related to the usage of a feather pillow are shown in Figure 2B and C, respectively. Corticosteroids were not administered to these patients because of patients’ refusal. Serum KL-6 concentrations in the patient with House-HP were significantly increased in the summer season and gradually decreased after environmental improvement and cleaning her house of fungi, but increased again in the next year when fungi regrew (Figure 2B). Serum KL-6 concentrations in the patient with Bird-HP were significantly increased in the winter but gradually decreased in the summer. She ceased coming to the hospital twice because her symptoms had improved naturally. She never used feather duvets, but a feather pillow that she used during a nap in the winter was finally proven to be the cause (Figure 2C).

In general, serum KL-6 concentrations are thought to be higher in acute HP compared to chronic HP. Kmax in patients with acute HP (n = 12) was significantly higher than that in chronic HP (n = 6) [2635 (1678) for acute HP vs 1036 (402) for chronic HP, P = 0.024]. Kmax-Kmin was also significantly higher in patients with acute HP than chronic HP (1883 (883) for acute HP vs 683 (81) for chronic HP, P = 0.005, Figure 2D). These results confirm that variations in serum KL-6 concentrations were significantly greater in acute HP than chronic HP. There is no difference in lymphocytosis between acute HP [48.0 (55.1) %] and chronic HP [18.4 (5.6) %], and no correlation between serum KL-6 concentrations at the same day of BAL and lymphocytosis in BAL fluid.

ROC curves were used to evaluate the diagnostic value of KL-6 concentrations or KL-6 ratios for the diagnosis of HP, House-HP or Bird-HP. Smax/Wmax used for the diagnosis of Bird-HP resulted in the largest area under the curve with 0.878 (Table 2, Figure 3), indicating moderate accuracy. When the Smax/Wmax cut-off level was set at less than 0.61 based on the highest Youden’s index, the sensitivity, specificity, diagnostic accuracy and likelihood ratio for the diagnosis of Bird-HP were 77.8%, 92.7%, 91.2% and 10.63, respectively. These results suggest that seasonal variation of serum KL-6 concentrations is useful for the differential diagnosis of Bird-HP.